The present disclosure relates to multi-level memory compression. More particularly, the present disclosure relates to storing pages of data in a memory at different compression levels based upon page utilization.
Computer systems compress data in order to reduce the amount of memory required to store data. Different compression algorithms compress data at different “compression ratios.” A compression ratio is the ratio between the size of data when the data is compressed compared with the size of the data when the data is uncompressed (e.g., original size). Another industry term used to compare the amount of compression that a compression algorithm achieves is “compression space savings,” which is computed by the formula “1−compression ratio.” As such, highly compressed data results in a low compression ratio and a high compression space savings.
Although selecting a compression algorithm to achieve a high compression space savings is typically preferred for memory utilization purposes, these “high” compression algorithms typically require an increased amount of time and processor resources to compress/decompress data when compared to “low” compression algorithms.